This briefing covers 20 cybersecurity and geopolitics
stories published around Tuesday, November 17, 2026,
including activity involving PLA, China,
and 3 disclosed vulnerabilities
(CVE-2026-124643, CVE-2026-124645, CVE-2026-124646).
Each entry links to the original reporting.
Mandiant tracked COLD RELIC — also known as Star Blizzard and Callisto — using single cloud account takeovers as launchpads for cascading third-party intrusions over the past year. The technique industrializes Russian espionage by converting one credential into persistent access across interconnected organizations.
Kaspersky identified a 2026 campaign deploying GitDoor and GraphDoor backdoors against government targets in Asia and Africa, routing command-and-control traffic through Microsoft SharePoint and GitHub to blend with legitimate traffic.
Google's Threat Intelligence Group attributed the first fully automated AI-driven cyberattack to Russian-speaking actors, noting the tools prioritize speed over stealth and leave noisy forensic traces.
Pakistani threat actors deployed fraudulent recruitment sites to surveil Indian defense sector job seekers, with secondary targeting of military personnel and contractors.
Analysis traces China's accelerating nuclear buildup to Second Artillery Corps doctrine shaped by the 1993 successful test of a continental-US-range ballistic missile, over leadership resistance.
Iran's IRGC launched drone strikes targeting a U.S. military radar system in Bahrain on November 17, 2026, marking a direct kinetic attack on American infrastructure in the Gulf. The strike represents a significant escalation from proxy harassment to overt IRGC action against U.S. installations.
The NSA ordered staggered arrivals at Fort Meade as technicians work to restore NSA Net, the agency's classified internal network. An outage on a classified signals-intelligence network raises immediate questions about operational continuity and whether the disruption is technical or adversary-induced.
A new U.S. rule mandates that overseas data centers serving China, Russia, Iran, and Cuba identify end users to block indirect access to American AI technology. The regulation closes a known loophole where adversary-linked entities routed AI compute through third-country cloud infrastructure.
Germany is spending a record €125 billion on defense in 2026, targeting Europe's strongest conventional army by 2039, but bureaucratic procurement processes are preventing capital from reaching capable startups proven in Ukraine.
China, Vietnam, and North Korea are providing Cuba not only diplomatic cover against U.S. pressure but exportable governance models for sustaining single-party rule under economic stress. The alignment deepens a bloc of authoritarian states sharing regime-resilience strategies across hemispheres.
The Iran conflict is accelerating a structural withdrawal of external great powers from Middle East regional order, leaving states to manage security without traditional patron guarantees. The power vacuum redraws alliance calculations for Gulf states, Israel, and Iran simultaneously.
A Just Security analysis applies game-theoretic modeling to argue that mutual restraint on frontier AI development is the dominant rational strategy for both Washington and Beijing. The framing attempts to recast AI competition as a coordination problem solvable by governance agreements rather than pure rivalry.
T. Elliot Gaiser, identified as the Trump administration's legal architect for justifying boat-strike operations in the Iran campaign, is under review for a senior nomination.
A supply chain attack on Optimizely injected malicious JavaScript into sites including CNN, Fox News, BBC, Salesforce, GitLab, and Zoom, targeting cryptocurrency wallets, payment data, and iOS device vulnerabilities.
Microsoft Threat Intelligence tracked a cross-sector phishing campaign using fake Microsoft 365 outage notifications to harvest credentials from organizations worldwide.
LLMs are enabling a broader pool of ransomware operators to generate convincing phishing content and rapidly identify high-value data, expanding the attacker base.
CVE-2026-124646, an insecure direct object reference vulnerability in India's e-Sushrut Hospital Management Information System, allows unauthenticated remote attackers to enumerate and access sensitive patient records via manipulated API parameters.
CVE-2026-124645 allows unauthenticated attackers to extract a static cryptographic key from e-Sushrut HMIS login pages and reproduce password hashes. Hardcoded secrets in a national health management platform expose patient data and administrative access at scale.
CVE-2026-124643 enables unauthenticated remote attackers to access e-Sushrut HMIS dashboards via crafted HTTP requests due to broken session validation. Chained with CVE-2026-124645, the two flaws effectively strip all access controls from India's hospital information system.
Threat actors hijacked Trivy, a widely deployed open-source security scanner, by injecting malicious code disguised as GitHub's official checkout action, bypassing SHA-pinning defenses. The compromise undermines a foundational supply chain security control trusted by enterprise CI/CD pipelines globally.